Vintage Year: Why the Year a Fund Closes Shapes Its Record

A fund's vintage is the year it is treated as having started, and providers date it differently: some from the first close, some from the first capital call, some from the final close. Nilgiri Growth Partners Fund II, invented, counts its own clock from its final close, four years after Nilgiri Growth Partners Fund I counts its. A vintage comparison that does not say which convention it used is comparing two different things.

One thing here is assumed by everybody and checked by almost nobody. When a fund is said to be of a certain vintage, it sounds like a fact about the fund, the way a lorry's registration year is a fact about the lorry. It is not. A vintage is a label somebody attached, using a rule they may not have written down, and the same fund can carry two different labels at the same time without anybody having made a mistake. A vintage is a statement about a fund made by a labeller, not a property the fund carries around inside itself.

Underneath the label, though, there is something entirely real, and it is what makes the label worth any attention at all. A fund's starting line fixes the stretch during which it went out and bought things, the stretch during which it had to sit and hold them, and the day by which its contract says it must be finished. Nobody inside the fund voted on any of those three. All three were set by a calendar. The calendar is what makes a vintage worth knowing, and the calendar is also exactly what makes a vintage so tempting to use as an explanation for outcomes it cannot explain.

What exactly is a vintage year, and is it a property of the fund?

A school labels its students in the same way. Somebody is from the batch that joined in a particular year, and the whole institution runs on that label. The batch label puts them in a section, it decides which examinations they sat and when, and thirty years later it decides which reunion invitation lands in their inbox. The awkward question is what the batch label says about that particular student. Almost nothing. The label says when they were in the building. The label does not say who topped the class, who scraped through, or who left in the second year and came back later. A batch label is a filing key, not an assessment.

A vintage yearThe year a fund is treated as having started, on a convention that must be stated. works in exactly that way. A vintage year is a filing key for funds. A private fund is a fixed-life thing with a beginning and a contracted end, so somebody keeping track of thousands of closed-end private vehicles needs a way of putting each one in a drawer, and the year it started is the most natural drawer available. Once every fund is in a drawer, the funds sharing a drawer form a cohortThe set of funds a provider treats as sharing a vintage., and everything that gets built on top of a vintage label is built on that grouping.

Two things follow immediately and both matter. The first is that the label depends entirely on where the labeller decided to draw the starting line, and three different days compete for that honour. The second is that a label describing a whole fund cannot say anything about the individual things inside that fund, in exactly the way a batch label says nothing about the individual student. The second point is where most misreadings of a vintage begin.

Why do three different days compete to be the same fund's starting line?

A private fund does not begin the way a shop begins, on the day the shutter goes up. A fund comes into existence across a stretch of months, and that stretch has at least three days in it with a genuine claim to being the start.

The first is the first closeThe point enough commitments are signed for the fund to begin., the point at which enough investors have signed for the fund to stop being an idea and start being a fund that can act. The second is the first capital callThe first notice requiring investors to pay in., the first notice that actually requires investors to send money. The first capital call has a claim too, being the first day anybody's cash is genuinely at work rather than merely promised. The third is the final closeThe day the offer shuts and this record starts each fund's clock., the day the offer shuts, the investor list is fixed and nobody else can join.

Each of those three is a defensible answer, and different providers of fund data have settled on different ones. The disagreement is not carelessness but three reasonable people disagreeing about what "started" means. The problem is not the disagreement. The problem is that a first close and a final close on the same fund can sit a year or more apart, so two providers applying two honest conventions can file the same fund under two different years and both be correct.

The claim being made is narrow. There is one fund. There is one run-up. Nothing about the fund changes along the strip. The only thing that changes is which of the three marked days was called the beginning. A turn of the calendar happens to fall between two of them, so that decision moves the fund into a different drawer. No fact about the fund did any work here at all.

What happens when the convention does not travel with the number?

Here is a small domestic version of the failure. Two households are comparing what they spend on groceries. One of them counts from the first of the month. The other counts from the day the salary lands, the twenty-eighth. Both are keeping honest records. Neither is lying about anything. But when they put the two totals side by side and one is higher, the higher number is partly an artefact of where each person put the boundary. Neither one wrote the boundary down next to the figure, so neither of them can see it.

The two boundaries are the whole of the vintage problem, scaled up. A conventionThe stated rule by which a figure or a label is produced. is the rule by which a figure or a label was produced, and a figure separated from its convention has quietly stopped being a measurement. Such a figure has become a number with a story attached that nobody can now check.

So the rule is simple, mechanical, and it is the single most useful thing in this guide: the convention travels with the number, every time, in the same line, in the same cell, in the same sentence. Not in a methodology note at the back. Not in an appendix somebody will not open. Beside the number, where anybody reading it later can see what it was built from. A vintage written into a note carries which day it was dated from in the same breath. Where that day cannot be established the note says so, and "convention not stated" is itself a finding.

What does a vintage genuinely fix, if it is only a label?

Now the useful half. Everything above says the label is soft. The thing sitting under the label is not soft at all, and that thing is the reason a vintage is worth any attention in the first place.

Every closed-end private fund carries two contracted stretches that both begin on the same day. One is the investment period, the years during which the fund is allowed to call capital and buy things. The other is the term, the total contracted life at the end of which the fund is supposed to be finished. Both of those run from the fund's own starting line. So the moment the starting line is fixed, three windows are fixed with it, and none of them was chosen by anybody in the fund.

The record worked through here is a single fund. Nilgiri Growth Partners Fund II, invented, has a five-year investment period and a ten-year term, both running from its own final close. The five-year period and the ten-year term produce the following, and every figure here belongs to that fund on its own clock.

The entry windowThe stretch during which a fund bought what it holds. is the one worth pausing on. The calendar does its most obvious work there. Nilgiri Growth Partners Fund II, invented, entered its first holding at its own Year 1 Q3 and its last at its own Year 5 Q3. Whatever conditions existed for buying unlisted businesses across those 4.00 years, on its own clock, are the conditions that fund met. The fund could not go back and buy earlier. Nor could it wait and buy later. After the end of its Year 5 its investment period closed, and capital could be called only for a short numbered list of purposes rather than for new companies.

Here is the everyday version. A household that finishes building a house has fixed, on the day of handover and without deciding to, when its structural warranty starts and when it runs out, which season it will be doing its first repainting in, and how old the building will be when the children finish school. None of that was a preference anybody expressed. All of it fell out of one date. A fund's vintage does the same job: it converts a single day into a set of deadlines that then govern behaviour for a decade.

How does this record date its own two funds?

The heading and the record worked through here do not use the same convention, and the disagreement between them is the lesson rather than a fault. The heading names the year a fund closes, and the most widely used convention in ordinary conversation is the first close. Both funds here are dated from the final close instead, the day each fund's own contractual clock starts ticking. The investment period and the term are both measured from that day.

So on this record every date is written as a year and a quarter counted from that fund's own final close. Nilgiri Growth Partners Fund II's Year 1 Q3 means three-quarters of a year after Fund II's final close, and nothing else. Two funds on this record therefore carry two entirely separate clocks, and a date that does not name its fund is not a date at all. That is not pedantry. Naming the fund is the only way two records can be laid next to each other without producing nonsense, and four years of offset leave a great deal of nonsense available.

How far apart are these two funds, and what does that look like on both clocks?

Nilgiri Alternatives Advisors Private Limited, invented, is the manager behind both funds worked through here. Nilgiri Growth Partners Fund I, invented, is wound up. Nilgiri Growth Partners Fund II, invented, is still running. Same manager. Same ten-year term. Same five-year investment period. Same management fee rate, same preferred return, same catch-up, same carried interest share, same whole-of-fund ordering. The documents are, for present purposes, twins.

One thing differs, and it is the whole subject: Fund I's final close falls four years before Fund II's, so Fund I's Year n is the same real afternoon as Fund II's Year n minus 4, and that offset holds for the entire life of both funds.

Every one of the three columns below is a sentence somebody could get wrong in a note, so read them slowly.

The fourth row is the one that breaks careless notes. The afternoon Fund I was wound up has two correct names. Fund I's Year 10 Q4 is right. Fund II's Year 6 Q4 is equally right. A bare Year 6 Q4 with no fund attached, handed to somebody six months later, is a sentence that cannot be reconciled by anybody, including the person who wrote it.

Two brothers each open a tea stall on the same street, one of them four years before the other. When the younger one opens his shutter on the first morning, the elder is four years into his own trading life, with regular customers, a known supplier and a shelf of things that did not work. Asking which stall is doing better on that first morning is not a question about tea at all. The question is really about which of them has been open longer, and the answer was fixed the day each of them opened.

What do the two records show, and why can they not be set against each other?

A private fund figure detached from its fund or its period is not a figure, so every figure in the two records below carries both.

Nilgiri Growth Partners Fund I, invented, finished its own contracted ten years at 1.80 times net to its investors and a net internal rate of return of 11.1 per cent across those ten years. Everything it had was sold. Everything it owed its investors was paid. There is nothing left in it to change.

Nilgiri Growth Partners Fund II, invented, stands at its own Year 9 Q2 record date at 1.50 times total value over money paid in, and a net internal rate of return of 8.3 per cent over those 8.50 years, with Rs 2,82,00,00,000 of that value sitting in five holdings that nobody has bought. Six quarters of its contracted term remain. Two points of that fund's curve are these: Fund II stood at 0.95 times and minus 1.8 per cent at its own Year 5 year end, and at 1.50 times and plus 8.3 per cent at its Year 9 Q2 record date. The full year-by-year shape of that curve is covered separately.

Now the important sentence, and it is a refusal rather than a conclusion. The two sets of figures are not comparable, so neither fund can be ranked against the other. The same manager ran both, so the manager cannot be ranked either. Three separate things stop the comparison. Fund I's number is a finished number made entirely of cash that changed hands. 39.2 per cent of Fund II's total value has never been sold to anybody and is a carrying estimate. Fund I's figure is stated net to its investors after everything the manager was contractually due. Fund II's is a total value over money paid in, and what sits between those two denominators is covered separately. And Fund I had run its full 10.00 years while Fund II has run 8.50 of its own, so one of them has had time to finish and the other has not.

The blunt statement is the one readers most often get backwards. Fund II is not behind. Fund II is younger. The two are different statements, and only one of them is supported by the record set out here.

What was Fund I actually showing on the day Fund II closed?

One fact about successor funds is uncomfortable and completely ordinary at the same time, and that combination is worth learning to recognise.

Nilgiri Growth Partners Fund II held its final close at a moment that sits four years along Nilgiri Growth Partners Fund I's own clock, at the beginning of Fund I's Year 5. Ask the obvious question: what did the earlier fund look like that afternoon, given that it was the record the later fund was raised against?

Fund I had made no distribution at all. Not a small one. None. Its first payment of any kind came at its own Year 6 Q2, more than a year after Fund II had closed and had already begun calling capital. Fund I had drawn Rs 2,20,00,00,000 of the Rs 2,40,00,00,000 it would ever draw across its whole life, being its first six calls. And its total value over money paid in, at that moment on its own clock, was below 1.00 times.

So Nilgiri Growth Partners Fund II was raised against a track record that was, on the day it closed, entirely unrealised. Every rupee of apparent progress in Fund I at that point was a carrying estimate on things nobody had yet bought.

The reflex here is the wrong one. An unrealised predecessor record is not an irregularity and not a warning sign, but the plain arithmetic of a ten-year vehicle raising its successor in its fifth year. A closed-end fund that waits until it has sold everything before raising the next one waits a decade, during which the team it built has nothing to do. Successor funds are therefore normally raised while the predecessor is mid-life, and mid-life is precisely when a predecessor's record is mostly estimates. The vintage gap and the unrealised record are the same structural fact seen from two sides.

The right thing to take from it is a question, not a verdict. The question is: how much of this record is cash that has actually moved, and how much is a mark on something still held? The question has an answer and the answer is checkable. "This looks bad" has neither.

What can a vintage not be credited with?

A fund-level label sits on top of a portfolio, and a portfolio is not one thing. A portfolio is a small number of separate businesses that had separate lives and produced wildly separate results. If the spread between those separate results is larger than the difference the label is being asked to explain, then the label cannot be carrying the explanation, and the arithmetic of that is not a matter of opinion.

So measure the spread. Here is what actually came out of each of these two invented funds, expressed as each holding's share of everything that fund realised.

Take the numbers one at a time, with the fund and the period attached to each. Of the Rs 4,38,00,00,000 that Nilgiri Growth Partners Fund II, invented, had realised to its own Year 9 Q2 record date, a single holding produced Rs 2,03,00,00,000, being 46.3 per cent of the whole. One holding out of nine produced nearly half. Two of its nine are carried below what they cost, and one of those two returned nothing at all and therefore has no width on the strip above. Of the Rs 4,80,00,00,000 that Nilgiri Growth Partners Fund I, invented, realised across its completed ten years, a single holding produced Rs 1,40,00,00,000, being 29.2 per cent of everything the fund realised, and it did that on 10.0 per cent of the money that fund invested. Two of its seven returned less than they cost. The unrounded shares on each strip sum to exactly 100.0, so nothing there has been nudged.

Now put the spread and the gap on one scale. Comparing their sizes any other way is not honest. Measured on cost, meaning each holding's total value divided by what that holding cost, Nilgiri Growth Partners Fund I stands at 2.40 times across its completed ten years and Nilgiri Growth Partners Fund II at 1.80 times to its own Year 9 Q2 record date. The gap between those two fund-level figures is 0.60 times.

Inside Nilgiri Growth Partners Fund I alone, across its completed ten years, the seven holdings run from 0.00 times to 7.00 times on cost, a spread more than eleven times as wide as the whole 0.60 times gap between the two funds. Inside Nilgiri Growth Partners Fund II, to its own Year 9 Q2 record date, the nine run from 0.00 times to 2.90 times, still nearly five times the gap. Whatever is producing outcomes at these two funds is operating at the level of individual businesses, and a label attached to a whole fund has no mechanism for reaching down there.

The household version is a shopkeeper with nine lines of stock. Eight of them tick along and one of them, for reasons specific to that one item, sells out three times over. At the end of the year the shop's overall margin is a single number, and that number is mostly the one item wearing a disguise. Ask why the shop did what it did and the honest answer is a sentence about that item, not a sentence about the year.

What does a vintage set that nobody inside the fund chose?

How Vintage Year Can Affect Private-Fund Context

The soft label and the hard windows are not in tension, and the honest position holds both at once.

A vintage genuinely reaches only context, meaning the circumstances a fund had to operate inside, and nobody in the fund selected those. Four of those are worth naming, and each one is a mechanism rather than a claim about outcomes.

First, it fixes when the buying had to happen. Nilgiri Growth Partners Fund II, invented, entered all nine of its holdings between its own Year 1 Q3 and its own Year 5 Q3. Whatever the conditions were for buying unlisted businesses across those 4.00 years, they were that fund's conditions, and a fund that closed four years earlier met a different set on its own clock. Neither manager selected which set they walked into.

Second, it fixes how long the fund can afford to be patient. The contracted term sets an outer edge, and the closer that edge gets, the smaller the range of things a manager can sensibly do with a holding. Nilgiri Growth Partners Fund II, invented, had six quarters of contracted term left at its own Year 9 Q2 record date and five holdings still held. Six quarters left is a fact about its clock and not a judgement about those five holdings.

Third, it fixes which conditions the selling has to meet. A fund whose contracted end arrives at one point on the calendar meets whatever conditions exist at that point, and a fund four years behind meets whatever exists four years later. Both edges were set the day each fund closed.

Fourth, it fixes who the fund is being compared with. Because providers sort funds into a cohort by vintage, the label decides which set a fund is placed alongside, and therefore what a reader sees next to it. How such a set is assembled, and what it needs behind it before any comparison is meaningful, is covered separately.

Notice what is absent from all four. There is no statement that any window was favourable or unfavourable, no statement that one fund's conditions were better than another's, and no suggestion anywhere that a reader should prefer one starting year over another or should spread anything across several. DispersionThe spread of outcomes between the holdings inside one fund. inside a portfolio, as the figures above show, dwarfs the fund-level differences a vintage could reach anyway. A vintage sets the circumstances. A vintage does not set the result.

What should be done with a vintage label?

Reading a vintage label reduces to one question with a stop attached, and the stop is the part most people leave out.

Ask which day the vintage was dated from. If the answer is the first close, the first capital call or the final close, write that answer next to the label and carry it everywhere the label goes. If there is no answer available, stop, and record that there is no answer available. A label whose convention nobody can state is not a weak measurement; it is not a measurement, and building a set or a comparison on top of it manufactures a difference out of a filing decision.

Then two follow-up questions, both mechanical. How much of the record attached to this label is cash that has actually moved, and how much is a carrying estimate on things still held? And what does the portfolio inside look like, holding by holding? The spread lives there, and the label cannot see it.

The whole procedure is three questions long. Three questions take a minute, and they are the difference between a note that can be reconciled by a stranger in two years and a note that quietly attributes a difference to a calendar.